– [Patrick] Joining us today for a special conversation is Dr. Robert Metcalfe. Many of you know him as the inventor of Ethernet, the communications backbone that really underlies almost everything that we do today, but you may also know him as the founder of 3Com and for Metcalfe’s Law. Bob began his pioneering work in 1970 at MIT followed by Harvard, Xerox’s Palo Alto Research Center, and Stanford. And we’re thrilled today to have him not just as an icon, but as a champion of networking to this day. For the last decade, Bob has been working with the next generation of technologist and business leaders as a professor of innovation at the Cockrell School of Engineering at the university of Texas in Austin. Bob is also the founding director of UT Austin Innovation Center. And he’s the professor of entrepreneurship at the McCombs School of Business and Murchison Fellow of Free Enterprise. More recently Bob became the principal investigator of Texas Geo, which is incubating startups to advanced technologies, delivering cheap, safe, clean, and ubiquitous geothermal energy. Bob, thank you so much for being with us today.
– [Robert] My pleasure. Thank you.
– [Patrick] So, here’s our first question, there have been a few of these research centers, cradles of innovation that were the birthplace of so many technologies that shape our lives to this day. And Xerox PARC was certainly one of those. We got laser printing, object-oriented programming, the modern mouse gooey, UI, and of course, Ethernet. So, what was it like to be there? Because it sounds amazing.
– [Robert] Well, it was heaven on earth. Imagine us tucked away in a building in the hills of Palo Alto where the temperature is always 71 degrees and the computer science lab had been formed by Bob Taylor who may he rest in peace, he was an alumnus of the university of Texas at Austin, by the way. But he then at ARPA, the Advanced Research Projects Agency of the Department of Defense was involved in funding, computer science around the world. And when it came time for him to build a computer science lab at the Xerox Research Center in Palo Alto, he picked a lot of people from the ARPA community, me included and then gave us a general mission, which was to research technologies related to the paperless office of the future, which is mildly ironic since Xerox’s entire business was based on producing paper by the ton.
And we joined in, we invented the laser printer and produced more paper than ever before, much to the frustration of the paperless office people. And it was like I said, heaven on earth. We had one meeting a week that we were obliged to attend, and the meeting was called Dealer. And I think it was noon on Tuesday, but I could be wrong about that. And so, we would have to… All of us were off writing computer programs and building things, but every once in a while, every Tuesday at noon we would come together, there would be a guest speaker. And then we would share progress reports, and everyone would lie about how much progress they made that week on their various projects, which was entertaining. And the speeches given at Dealer were recruiting speeches.
People we were recruiting come and give a presentation at Dealer, and that would be part of the recruiting process. So, Palo Alto, endless summer, no obligation to do anything. I didn’t have to raise money; I didn’t have to teach courses. Now, Xerox PARC could be viewed, should be viewed I think loosely as a lab of Stanford university. So, the lab was put in Palo Alto to be near Stanford, and then most of the researchers became professors at Stanford. So, I became a consulting associate professor at Stanford engaged with some of the research there, but also mostly teaching a course every semester. So that was fun too.
– [Patrick] That sounds amazing. The idea of pure R&D really does sound like heaven on earth, certainly for a lot of our audience who they don’t necessarily get to do that, and they have more than one meeting a week. What’s it been like since then to watch so many of those technologies really just become taken for granted, they are ubiquitous and everywhere? Did they sort of go off on their own or do you still get a feeling of ownership for a lot of that work?
– [Robert] Well as you’ve indicated, a lot of really cool things that we take for granted came out of various labs in the computer science lab, but a problem developed in that there was no way for those technologies to be applied anywhere. Xerox, the company itself was slow to adopt these technologies and certainly slow to turn them into products. So then what started happening is people started leaving PARC to start companies, I left to start 3Com, Adobe spun out, the list goes on, lots of companies. And then Xerox said, “Wait a minute, we’re leaving money on the table here.” So, Xerox started investing in some of these companies as they spun out. And so, there’s a lot of lessons and books have been written about the innovation process.
One of the books is named Fumbling the Future, and it’s about how Xerox failed to exploit the technologies which it had developed in its lab. I think the books a little rough on Xerox. One thing my colleagues do at PARC when they talk about the good old days at PARC, they refer to Xerox as “they” and I hasten to remind them that no, it was us. We were making the mistakes, it’s not really fair to blame Xerox for everything. Plus. Xerox did actually get a lot out of PARC in particular, the laser printer business became a multi-billion–dollar business spun out of the Xerox Research Center.
– [Patrick] So, you mentioned that technology, or at least the way that innovation happens, where it happens, the mechanisms have changed over the years. So how has that changed from that pure R&D corporation back to where it’s coming from now?
– [Robert] Well, the big change you’ll hear as a lament. People will refer nostalgically to the good old days of Xerox PARC or the good old days of Bell Labs, which was the father of all these or mother of all these research labs. And they’re pretty much gone now, Bell Labs still exists, but it may have 1000 people instead of 25,000 people. Xerox PARC still exist in the hills of Palo Alto, but its emphasis has shifted and you don’t hear so much about it. It’s owned by Xerox, but half of its work is for Xerox and the other half is contract work for sponsors. So, there’s this funny thing, oddly I graduated from the last good class to graduate from my high school. And then I graduated from the last good class to graduate from MIT. And then I graduated from Xerox PARC the last year of its preeminence as a research center. Why is it that I always see my departure as the end of wherever I’m leaving? Oh, I know what that is, that’s called ego, uncontrollable ego. That’s what it is.
– [Patrick] Well, okay, so here’s one for you. You’re in the Tuesday noon meeting, you get to have some amazing pitches. People are talking about technology and obviously not every technology spun out. So, what was a project that you saw that would have maybe changed the world that never made it out?
– [Robert]There’s quite a diversity of projects all the way from the network, all the way up to the… But databases were not on the scheme. The World Wide Web was not anticipated by this research. Those are two gaps. Mobility was not, these were desktop workstations we were working with. So, we had taken the step from timeshared minicomputers, to desktop PCs with local area networks, connecting them, but we didn’t make the step to the World Wide Web, at least not while I was there. And we didn’t make the step to mobility, both of which are very big today and were not anticipated by the PARC research.
– [Patrick] So, connectivity really ended up being a critical part of that, because it’s hard to imagine things like modern mobility based on what expectations for bandwidth would have looked like in the 70s?
– [Robert] Yeah, well, bandwidth was scarce until about the early 70s. Ethernet, for example, the day before I had Ethernet installed in my office, I had the latest computer terminal Texas Instrument Silent 700, it ran at 30 characters per second, and the RS-232 network carried bits at 300 bits per second. The next day, the Ethernet gets installed at 2.94 megabits per second. And I’ll do the arithmetic for you, that’s an increase by a factor of 10,000. So, the bandwidth, I think Ethernet sort of was part of the revolution that said bandwidth is not going to be scarce, bandwidth is not going to be abundant. And by abundant, eventually that meant uploading cat pictures, prior to Ethernet, you would never upload a cat picture. It would take too long and be too expensive, but now you can do it anytime you want.
– [Patrick] So that was a real part of the focus on Ethernet was that it was going to be able, or it was the solution to be able to deliver the bandwidth, that was otherwise throttling a lot of innovation around it.
– [Robert] That’s right. And so, for example, we were building this laser printer, it was page per second, 500 dots per inch. So, 8.5 by 11 by 500 by 500 per second is 20 megabits per second. So, the old RS-232 interface running at the most 10 kilobits per second would never keep that printer busy. The printer would be sitting there downloading files all day and not printing, so that was one of the requirements that pushed Ethernet to be fast. So, one of the principles associated with the Ethernet brand is build it, and they will come. Build the Ethernet, the current Ethernet as fast as you possibly can with existing semiconductors and then the applications will follow somehow serendipitously. So, Ethernet started at 2.94 megabits per second, but today they’re beginning to get ready to ship 400 gigabits per second. So, there’s been a huge increase as the semiconductors got better, the networks got faster.
– [Patrick] Yeah, it’s been amazing all those high speeds where it goes back and forth between fiber and then copper continues to be able to push bandwidth, that really would have been hard to imagine. Was it pretty straightforward getting it off the ground or you ran into barriers or what was that process for really helping it take hold?
– [Robert] Well, I see it as a battle. One of the things I teach is the difference between what I call the invention and innovation. And invention is done in the lab and everyone loves you and you’re sponsored and you’re a celebrity inventing stuff. And then you try to innovate and innovation that is bring it to market and have it adopted by the world markets. And that is not a nice process. The status quo is resourceful and nasty. So, Ethernet won a war, we call it the LAN wars, the Local Area Networking Wars that Ethernet eventually won, but it took 20, 30 years for Ethernet to win that battle. And the struggle started right at Xerox Research. David Boggs and I built the first Ethernet. We hadn’t even built it yet when a physicist at PARC sent a memo to my boss’s boss saying that Ethernet, by the way, without talking with me, which means he was a physicist, he didn’t understand people.
So instead of talking to me about what we were doing, he just fires off a memo to my boss’s boss saying Ethernet was basically a fraudulent project because it was not quantum noise limited. That is that cable that we were sharing to send our packets was not draining every bit per second it could out of the cable, but of course to do so would’ve cost money to build the electronics and so on. So, we ran at 2.94 megabits per second, and he tried to kill our project because it wasn’t quantum noise limited, which is of course not one of our goals. So that was just the beginning of the fights. And it lasted when we tried to standardize Ethernet, which we did through the IEEE starting in 1980, IBM and General Motors, two large companies decided they didn’t like the idea that Xerox would make a standard called Ethernet.
So they developed their own alternatives. IBM introduced the IBM token ring and General Motors introduced the General Motors token bus and they both showed up at the IEEE to derail Ethernet, which was in the process of being standardized. It got so ugly that the IEEE courageously standardized all three. So instead of choosing one, the IEEE chickened out and made the IBM token ring, and the General Motors token bus and Ethernet all became standards. IEEE 802.3 was Ethernet; IEEE 802.4 was token bus and IEEE 802.5 was IBM token ring. And it took a decade or two to kill them because they, well, IBM was the dominant computer company in the world, General Motors was General Motors. I remember telling the guy, “Hey, why don’t you guys make cars and I’ll make the networks?” And they said, “No, we’re at the customer, and we have a particular idea, the token bus.” That one was easy to kill that died in a single digit number of years, the token ring on the other hand, persisted for decades because of the preeminence of IBM.
– [Patrick] Yeah. My first enterprise gig, it was token ring, NetWare and token ring and it was a thing. So-
– [Robert] My two launch enemies, NetWare, so Novell and 3Com, my company were competitors for about three or four years. We were at each other’s throats. And token ring lasted way longer than that because of once again, IBM’s preeminence.
– [Patrick] It was a good day when we finally switched out the big brick four pin connectors from token ring to a nice RJ45. So, IEEE standardizes all three standards. So, then it’s the technology just one, or it turned into a sales process, or what was the next step in terms of market adoption?
– [Robert] The token bus died quickly, so the token ring became the focus. And basically, everybody except IBM was in favor of Ethernet and against IBM. And so, we had HP, we had Deck on the Ethernet side plus we had a flock of startups, including 3Com, but many startups, and the Ethernet patents were put into the public domain as part of the standardization process. So, startups were cropping up all over the place, competing with 3Com, for example, and IBM didn’t really understand the game. They saw the standardization of Ethernet, and so they said, “Okay, we’re get a standardized token ring, haha,” but their heart wasn’t in it.
In their heart of hearts, they were used to setting standards. So, 3Com shipped IBM token ring ahead of IBM, startups are generally faster than big companies, but then we had trouble selling it because IBM had software dust that it sprinkled over it’s hardware. So, we had a hard time getting our token ring cards to work in IBM customer sites. So, IBM maintained dominant market share in token ring, whereas Ethernet, it was hard to maintain a dominant market share there because it was so competitive, fiercely competitive among the Ethernet suppliers.
– [Patrick] But you started talking about the value of networking as a part of 3Com’s message. To be able to enable innovation and technology, you needed to actually be able to get the value of the network and this is where Metcalfe’s law came from. So, talk to us about the formation of that, and whether you consider that as sort of sales and adoption innovation, or is that still technology engineering innovation?
– [Robert] Well, we developed an Ethernet card for the IBM PC and started selling it. And we learned that no one had any IBM PCs, they were brand new, so it was very hard to find customers. And then they were reluctant, they were just learning about PCs, so they were reluctant to do adventurous things. So, we offered a three-node trial network for $3,000. You got three of our cards, boom, boom, boom, and the three PCs and a cable to connect the three PCs together and a diskette with software on it. And what that software allowed you to do is three important things, share a printer, share a big disk, and exchange emails among three PCs. The printers were expensive, like the laser printer from Apple was a $7,000 printer. So, the idea of sharing it made sense, today printers cost $100, it’s a different scene.
IBM had just come out with a 20–megabyte hard disk, which no one knew how to use, it was just so huge. And we allowed you to have three PCs, all keep their files on a shared IBM disc. And then there was exchanging emails. So, our customers bought these trial kits from this little startup out there in Santa Clara, I guess we were in Mountain View at that time. And they worked, you could share a printer, you could share a disk, you could exchange email, but the customers all said, “That’s fine and good. And you kept your promise, they do what they do, but it’s not useful.” Well, that’s a terrible thing to hear from your customers. And I was head of sales and marketing at that time, so it was my job to solve this problem.
So, I went to Sanford one night and went into a trance and developed a 35-millimeter slide. We didn’t have PowerPoint, understand, that showed that the cost of a network was sort of linear in the number of my cards that you bought, but the value was the number of potential connections. And the number of potential connections among N nodes is approximately N squared. So, then I associated that with value that it was valuable to connect things together hence the formula, V grows as N squared, the value of a network grows as the square of the number of attached people, devices, whatever. And this slide was not called Metcalfe’s Law, it was just a 35-millimeter slide. I made six copies of the slide because my salesforce had six people in it.
And they carefully inserted this slide into their carousels with their other 35-millimeter slides and went out to the customers and basically told the customers, “The reason your networks are not useful is they are too small, and the remedy for that is to buy more of our products,” and they did. And so, with that pitch alone, we went public in 1984, based on the sales stemming from that. And it paid off. That is, when they grew the networks, the networks actually were more valuable. Then the question arises when I made this slide up and I gave it to the salesforce, was I lying? I was after all a sales and marketing executive, was I lying when I made that slide? And the answer is no, because I had the benefit of a time machine.
I was able to go into the future with a time machine. And that time machine was the Xerox Palo Alto Research Center, where I had enjoyed the… What Dave Boggs and I and others did there is we filled Xerox with PCs and LANs and routers and printers; we basically built a prototype of the internet inside of Xerox. And then when it came time to start the company, I came back from the future in my time machine. And when I wrote that slide, I knew that it was correct. I knew because I’d seen it, we all saw the inside internet at Xerox and saw that it was good. So no, I was not lying when I made that slide. Now that slide in 1995, 15 years later was picked up by a guy named George Gilder, a famous futurist. And he called that slide Metcalfe’s Law, so I’ve been defending it ever since.
– [Patrick] So, there’s a question there then, and this goes back to innovation. You started as a practical engineer working on the protocol and you end up as a sales and marketing executive engineering and equation to help demonstrate that while you’re limited to three cards in your starter kit, we have way more nodes running in our lab, and this is how you can actually measure the value. So, is that a career change or is that just an extension of technology innovation to help move something forward and drive adoption?
– [Robert] Well, for me, it was a career change and a very unusual one. It’s very difficult to convert an engineer into a marketing person, by the way, it’s much harder to convert a marketing person into an engineering person. So as hard as both of those are, I did the easier one, but very few people choose to make such a transition and I did. But that’s not a determinative of the innovation process, what we were using was a time-tested model of innovation called a Silicon Valley Startup System. And we just fell into it and leveraged, we had venture capitalists, we had the titles on our business cards were all borrowed from HP. We had program managers, project managers, senior engineers, all titles that had been finally developed and specified by HP, we just adopted them full speed. So, there’s my making the transition from engineer to salesman, and then separately, there’s 3Com using the Silicon Valley Innovation System to put Ethernet into business, into the internet.
– [Patrick] Well, the question is a little bit personal for me as well, because most of my career, I was an engineer and now I’m not marketing exactly, but I play with marketing. So, a lot of that is that thought about remembering that business is about making money, but for engineers or a lot of engineers that are maybe in IT and in operations work is really about innovating to solve problems. You’ve been on both sides of that though, so how do you encourage tech professionals to maybe either ride that line between leadership and engineering or sort of innovation and business development versus profitability?
– [Robert] Well, one of the problems as head of sales and marketing is that the sales–people in our company were paid more than engineers. That is, they had quotas and so their compensation was contingent on successful selling, but they had a lot of upside and generally salespeople were paid more than engineers. And this annoyed my people, my engineers were annoyed by this. So, I did the, I’ll share this secret this is what you do when you have this problem, I said, “Okay, you can be a salesperson now.” I said this to the engineer, and we’re going to Syracuse tomorrow, and we’re going to take the red eye and then we’re going to wake up for a 7 a.m. breakfast meeting with a bunch of hostile people in Syracuse, New York. No, Schenectady, I’m sorry, where the GE center is I remember.
We’re going to stay at a Ramada Inn, I’ve stayed at every Ramada Inn that there is, I think. And we’re going to answer hostile questions over breakfast tomorrow, and then we’re going to go from there to mid-morning and then we’re going to go there for lunch. And then we’re going to suffer fools gladly all afternoon, because our competitor has gotten there ahead of us and has poisoned the water with all sorts of accusations about our products, which we’re going to have. So, the engineer after about a week of this would see what was up and go back to engineering, which was… So, engineers and salespeople are both carbon-based life forms, but they’re substantially different. And I’m not sure I’m answering your question, but now it’s not as if engineers are not interested in money, there you have a case where it annoyed them, that salespeople were paid more than they were, and then there’s the question everyone wants to advance.
And if you’re an engineer, do you advance by being a more of an engineer or more of a manager? So, we at our company and other companies too developed a dual track where you could get promoted and increased compensation on either track, either as a manager or as an engineer, you didn’t have to switch to management in order to make more money. We wanted to eliminate that. And while we’re speaking of money, we discovered at one point that the development engineers at our company looked down on the manufacturing engineers at our company.
And so, there was a status, an annoying cultural status thing between development, engineers, and manufacturing engineers, but we were shipping quantity, million products at one point. And those manufacturing engineers just by tweaking a little something good, make an extra $10 billion of profit that year by speeding up some process. So, they were very important to our company at quantity of million. And so, one day we just went around and gave all the manufacturing engineers a 20% raise, just boom and made public the idea that those engineers were as important, if not more so than the development engineers, that was a cultural problem.
– [Patrick] Yeah, from my experience, I’ve been in software for most of my life and I have a deep respect for anyone in semiconductor and hardware, because you can’t fix it after it’s deployed. The ramp up time from engineering to production is pretty long. So, it’s got to be actually engineered, not just sort of delivered at the time that it appears on the shelf. So let me take that back a little bit and ask another question. So, what was the biggest surprise over the last 20 years as we’ve moved from central, well-funded pure R&D research centers, and we’ve seen folks transition between engineering and sales and marketing and advocacy. What was the biggest surprise to you in tech and what really made that stand out?
– [Robert] Well, there’s a lot of surprises, so it’s hard to pick one, but for example, in the mid-90s there emerged a thing called dense wave division multiplexing of optical fibers, something I know nothing about. It was a silver bullet for the internet, and it was a surprise because suddenly the cost of sending bits across country went down by factors of 10, thanks to dense wave division multiplexing, something that no one had ever heard of before, it came out of Bell Labs and then was picked up by startups and suddenly bandwidth was cheap. You just had to bury these fibers here and there, or string them from a pole, that was a surprise. Another surprise is when this all started, we thought the internet was going to connect these timeshared minicomputers with interactive dumb terminals, teletypes and glass teletypes, and 3270s of IBM.
And the PCs came along and suddenly we were not building the internet for interactive timesharing, we were building the internet for desktop personal computers. Ethernet, for example, took a big step forward until ethernet, the packet switching of the internet stopped at the interactive timesharing machine. And then its software talked to these dumb terminals and typed characters on them. Ethernet carried those packets, the Internet’s packets all the way to the desktop. So, software running on your PC could use the packet nature of the internet directly, it didn’t have to go through a dumb terminal emulation. So that was a big surprise. Another big surprise is at the beginning of this, this internet thing, this 3Com thing, it was nerds selling the nerds. I mean, I was head of sales and marketing, and that worked because my customers were fellow engineers mostly.
So I had standing, and I could talk their language. The internet changed all that starting in the mid-90s with the World Wide Web, suddenly the important people on the internet shifted, the nerds became less important, and the channel marketing people became more important and all the numbers started having zeros after them. So, my company, when we went public in March of 1984, we raised $11 million, which is not what you do anymore because the markets are so much bigger. Our markets were a hundred–million–dollar–a–year company was a big deal. And now you have to be a hundred–billion–dollar–a–year company to be important. So that was a surprise the shift, the scaling up of information technology in general and the change of people necessary to be successful change. And that was a surprise also.
– [Patrick] So, you mentioned nerds selling to nerds. And that certainly involves a lot of buzzwords and almost selling innovation as a quality in and of itself. But that hasn’t exactly changed if you’re on the ground in an IT organization, vendors come in and they hit you with a lot of buzzwords and a lot of promises like right now there’s a lot of focus on AI and ML. How do individual technologist, how do IT professionals navigate that sort of gig–focused feeds and speeds buzzwords to actually glean the technology that will be really useful?
– [Robert] Well, now you’re talking about my people, you’re talking about my customers, 3Com sold to network people, network managers, IT managers, IT professionals in general. And I was one of the vendors showing up with speeds and feeds, Ethernet and networking in general is and was fiercely competitive. And so, I’m really good at speeds and feeds and why my products are better than the other ones and so on, and I have a lot of respect for the… One of the things that my people did is make standards, they would choose, I remember when I was running InfoWorld, the IT department came forward and said, “There are people using Macintoshes here and we’d like to stop that because we’d like to support only Windows machines.”
And I said, “Well, you’re never going to get a Macintosh away from the art department.” So, we had most magazines in art department where everybody wears black all day, and those people were never going to give up their Macintoshes. So, there’s that corporate standard thing that we were always running into. So, when I’m out peddling Ethernet to CIO’s the high end of the IT chain, we had no success at all. In our day CIOs were all effectively employees of the IBM corporation. And the expression that we used at 3Com was that at that account, we’re going to have to wait for the CIO to die, because we’re never going to sell to that guy because he’s an IBM guy through and through. So, what we at our company had to do was sell to people other than IBM CIOs.
So we went down to work groups and began selling to people, to network managers who were much closer to the business than the corporate standard staff. And we slipped in and we came in through retail and through distributors instead of through a direct Salesforce like IBM. So, we just dodged all the IBM culture out there and just came in underneath it. So, then the next important question is why would a sane IT manager buy products from a startup? Why would you buy products from a startup if you’re an IT guy? And I know what the answer is, because I run into it time and time again, our customers had competitors and they wanted our products and we’re willing to risk go out on the bleeding edge of technology in order to gain advantage over their competitors. So, in essence, I have a great deal of regard for a competition, competition makes the world go round. So, we were competing, but our customers were also competing. And that’s what opened them up to buying products from the dirt ball little startup like we had.
– [Patrick] Was that a little bit like the relationship between artists and art critics when you’re talking about sort of frontline engineers and then leaderships where in art, a lot of the not so much market, but world of art and its promotion is led by art critics, but the engineers can sometimes feel, or the artist sometimes feel like their creativity is a little bit constrained by that system. How can IT professionals and managers help sort of break out or maybe communicate better when confronted with a real buzzwordy technology investigation or a chance to focus on the business value of technology instead of just, “Hey, this is cool, let’s implement it.”
– [Robert] So, where I come out on that is that I don’t view buzzwords pejoratively. That is, I think buzzwords play a role in the curation of new technologies. And we use buzzwords to navigate, to get the sense of very complicated things. So, I pay attention to language and think buzzwords are a tool of innovation rather than just marketing bulls*** that people make. Although there’s a lot of that, I net it out that IT managers need and should use buzzwords as a tool for discussing what’s going on in the space.
– [Patrick] So, they can actually be a center for conversation in a way of de-jargonifying the conversation between different groups.
– [Robert] It’s also shorthand and it’s also branding and it’s a mix of things. So, the PC LANs, have you heard that expression?
– [Patrick] Nope.
– [Robert] So, you’re too young. So, the buzzword in the 80s during our big move was personal computers arrived. And for a while, Bill and Steve wanted their PCs to be standalone, but then Ethernet came along and other LANs and then it became PC LANs as a buzzword. Now I’m talking buzzword here, personal computers connected by local area networks. And then after that, that evolved into client server where, have you heard of that term client server?
– [Patrick] Oh, yes.
– [Robert] Okay. That’s a major buzzword too. And it followed on PC LANs, and it described the process where you actually added servers to the network like database servers or routers, and then you access them over the LAN. So, there’s a succession of buzzwords, minicomputers, time sharing PCs, PC LANs, client server. And now what is it called, now cloud, now it’s the cloud computing.
– [Patrick] It was web and mobility and cloud, yeah. And now artificial intelligence and machine learning.
– [Robert] Yeah. Now that’s a special case. AI is… My thesis advisor in college was Marvin Minsky, one of the founders of AI. So, I’ve been watching AI for 50 years and it comes and it goes, and so this is like the third or fourth or fifth time AI and machine learning has captured the limelight. This time I’m more optimistic because my analysis of past failures of AI, and that’s too strong, but fading of AI had to do the AI depends on data and they would run out of data and then it was sort of plop, but now we have the internet, and the internet can deliver data. So, I’m much more optimistic about AI this time because the internet is there and it’s there to feed data to the AIs and the AIs without data they flop. But with data, they do amazing things.